Air Interface Challenges and Solutions for future 6G

Networks
Benoit Miscopein, Jean-Baptiste Doré, Emilio Strinati, Dimitri Kténas, Sergio
Barbarossa

To cite this version:

Benoit Miscopein, Jean-Baptiste Doré, Emilio Strinati, Dimitri Kténas, Sergio Barbarossa. Air Inter-
face Challenges and Solutions for future 6G Networks. 2019. �cea-01986524v2�

HAL Id: cea-01986524

https://hal-cea.archives-ouvertes.fr/cea-01986524v2
Submitted on 23 Jan 2019

HAL is a multi-disciplinary open access L’archive ouverte pluridisciplinaire HAL, est

archive for the deposit and dissemination of sci-
entific research documents, whether they are pub-
lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.
destinée au dépôt et à la diffusion de documents
scientifiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.
Air Interface Challenges and Solutions for future 6G
Networks
Jean-Baptiste D OR É1 , Emilio C ALVANESE S TRINATI1 , Benoı̂t M ISCOPEIN1 , Dimitri K T ÉNAS1 , Sergio BARBAROSSA2
1 CEA-Leti Minatec, 17 rue des Martyrs, 38054 Grenoble Cedex 9, France
2 Sapienza University of Rome, Italy
{emilio.calvanese-strinati}{dimitri.ktenas}@cea.fr, sergio.barbarossa@uniroma1.it
Abstract—5G networks are expected to be deployed in 2020
and are considered as a global game changer from a technological,
economic, societal and environmental perspective with very ag-
gressive performance levels in terms of latency, energy efficiency,
wireless broadband capacity, elasticity, etc. Many experts say that
the next big step for cellular networks is not 5G but its cloudifica-
tion that will support the explosion of radically new services and
applications ranging from immersive five-sense media to ambient
sensing intelligence and a pervasive introduction of artificial
intelligence. In our vision, the next generation of wireless systems
will transform the 5G service-oriented networks into user and
machine ad-hoc dynamic (re)configuration of network slices. This
will be enabled by software-defined end-to-end solutions from
the core to the radio access network, including the air interface
as well as the RF and antenna systems which are envisioned
as one of the keys to meet the user/service requirements. Users
and machines will be indeed able to dynamically (re)configure
network slices thanks to intelligent personal edges. This paper
presents our perspective of the 6G air interface and raise the
concept of software defined artificial intelligence and air interface
(SD-AI2 ) as a framework of 6G air interface. This concept is an
extension of the one initially proposed for 5G [1]. Instead of a
global optimized air interface, we envisage to bring agility and
flexibility to air interface with the help of artificial intelligence and
learning techniques to improve efficiency. The paper describes
the proposed context and highlights the technical challenges at
different levels.
Keywords—6G, beyond 5G, Mobile Edge Cloud, Artificial
Intelligence, Air interface.
I. I NTRODUCTION
In 2018, intensive successful testing, proof-of-concepts and
trials [2] have supported the launch in 2019 of the fifth
generation (5G) services. Today, even though 5G is not in
place yet, sixth generation (6G) research has already kicked-
off all around the world. Academic, industrial and research
communities are working on the identification of relevant
key enabling technologies that might define the so called
’beyond 5G’ (B5G) [3] or 6G. As an example, the authors
of [4] discuss on what should there be in 6G that is not in
5G or in its long term evolution. The paper also describes
potentials of new communication infrastructures incorporating
new distributed architectures, ubiquitous super 3D connectiv-
ity, pervasive introduction of distributed artificial intelligence
mechanisms. Semantic and machine learning tools at the edge
of the network, new human-human and human-machine in-
teraction mechanisms including 5-sense media and interactive
communications are also some envisaged concepts for the next
generation.
Among several collaborative research activities that are
focusing on 6G, on can cite the Comsenter project funded
by DARPA. This project aims to explore the potential of
new spectrum frontiers, in the range of sub-TeraHertz (sub-
THz). The objective is to design base stations working in
the sub-THz spectrum able to transmit up to 10 gigabits per
second (Gbps) per beam with a target of 10 terabits per second
(Tbps). The definition of the air interface with adaptive massive
spatial multiplexing optimized of sub-THz RF components
constraints and combined with cm-precision localization, sup-
plementing GPS, as well as the use of imaging techniques
to locate communications partners are the new challenges to
address. The 6Genesys flaghship initiative has also kicked-off
research on 6G focusing on several challenging research areas
including reliable near-instant unlimited wireless connectivity,
distributed computing and intelligence, as well as materials and
antennas to be utilized in future circuits and devices. China
has also started working on future 6G networks foreseen for
commercial services by 2030. Moreover, in July 2018 a new
ITU-T Focus Group Technologies for Network 2030 (FG NET-
2030) was established by ITU-T Study Group 13. Its goal is to
investigate the requirements, use cases, capabilities, and future
architecture of the networks for the year 2030 and beyond. In
addition to the above mentioned initiatives, other countries and
regions are also carrying out 6G conceptual design and R&D
work.
From an air-interface perspective, the 3GPP fifth Gener-
ation New Radio (5G-NR) standard proposes to adapt the
air interface configuration to the provided service and carrier
frequency. By doing so, the requirements for each service
can be satisfied. However the air interface is usually statically
configured after field trials run by operators before commercial
deployments. Moreover, it has been demonstrated that the
coexistence of different air interface configurations within
the same spectrum chunk can be ensured only with a huge
compromise in terms of spectral efficiency [5]. On the other
hand, the concept of network slicing is expected to play a
critical role in 5G networks because of the very wide gamut
of expected use cases and new services. Slicing is currently
foreseen as a service-level. Extending this concept down to
the physical layer remains an open question. To address these
challenges, we propose the concept of software defined artifi-
cial intelligence and air interface (SD-AI2 ) as the framework of
the beyond 5G air interface. This concept, mainly inspired by
SD-AI [1], provides a scalable and configurable mechanism to
customize the air interface design and to support user-centric
services based on the potential of latest artificial intelligence
(AI) breakthroughs.
Moreover, at millimeter and sub-THz bands, the commu-
nication links may become highly unpredictable and poten-
tially inefficient because of blockage events. To cope with
the intermittent nature of millimeter and sub-THz spectrum
channels, evolved multi-link mechanisms [6] and software
defined adaptive air interface will be needed to guarantee an
almost deterministic delivery of delay-sensitive applications
over such intermittent random channels.
II. AI- BASED S OFTWARE - DEFINED AIR INTERFACE
In 3GPP release 15, 5G-NR defines two frequency ranges
and a set of waveform configurations. Changing the nu-
merology allows the adaptation of symbol duration and the
SubCarrier Spacing (SCS) to the propagation conditions and
the service to be provided. For instance, enlarging the SCS
increases the robustness against Doppler shifts occurring in
mobility scenarios and strong phase noise of millimeter wave
frequencies. In parallel, reducing the symbol duration is bene-
ficial to reduce latency. We can easily assume that the 5G-NR
numerology system will be upgraded in the future to support
new bandwidth such as sub-THz. One can envisage for the
future 6G air interface to define mechanisms to dynamically
adapt the numerology of one link to the environment (presence
of LOS, Doppler Shift, delay spread of the channel). To this
end, multiplexing the numerologies should be performed in an
efficient way and with a high spectral efficiency.
The development of software-defined radio techniques of-
fers opportunities for the implementation of smart, flexible and
configurable air interface. The main evolution of the proposed
concept with respect to SD-AI from [1], is the possibility to
dynamically adapt the waveform, the link adaption strategies,
channel and bandwidth based on the context of a user (ge-
ographical position, motion, services, etc.). Many challenges
have to be addressed ; three of them are listed hereafter.
First, the design of an efficient waveform supporting the
multiplexing of various waveform types with a limited self-
interference generation needs to be considered. It has been
demonstrated in [5] that the current 5G-NR waveform, based
on OFDM, is not suitable for multiplexing different numerolo-
gies (i.e different SCS within the same band) as resources are
lost by the required introduction of guard band. The use of
filtered waveforms, as proposed in [7][8] could give an efficient
way to limited the inter numerology interference. Therefore,
one can envisaged to dynamically change the structure of the
waveform (SCS, guard interval, pilot distribution) according to
the channel conditions. Secondly, the use of the spatial filtering
by configurable antenna radiation patterns is also an efficient
manner to limit the mixed numerology interference by isolating
numerologies in different radiation beams. At high frequency,
antenna integration allows efficient multi-user beamforming
schemes, but when it comes to sub 6-GHz bands, the design
and integration of small and energy efficient antenna systems
remains an open challenge. Current works are focusing on
hybrid beamforming schemes, i.e mixed digital and analog [9].
Hybrid beamforming transceivers combine high dimensional
analog phase shifters and power amplifiers with digital signal
processing units. This hybrid beamforming design reduces the
cost and power consumption but require complex optimiza-
tion problem to solve. Second, With the growing interest of
artificial intelligence (AI) technologies in image and signal
processing, the introduction of AI technologies into wireless
communication systems could introduce a new paradigm in
the design of modern wireless system. Currently, many re-
searches are focused on the use of neural networks for solving
difficult problems, such resource allocation optimization in
wireless communication systems. One can envisaged to go
forward by implementation neural network to solve hybrid
beamformer optimization problems, see [10] as an example. In
the framework of our proposed SD-AI2 , it could be considered
to collect and analyze channel state information (CSI) as well
as localization context (including RSSI, GNNS signal etc.)
with a learning network to classify and select the best set
of parameters to exchange data, including channel, frequency
and spatial resources to be used as well as numerology and
modulation and coding scheme. Some preliminary works have
been done in the identification and classification of propaga-
tion channel condition based on machine learning [11][12].
Third, the virtualization and cloudification of the complete
RAN including air interface as well as antenna system will
introduce new challenges at different levels. The need of
high throughput low latency (possibly wireless) links for the
backhauling system and the increase of computation load
needed will be a challenge researchers and engineers will have
to solve. Distributed computing and caching techniques would
probability will play a key role in solving the unprecedented
requirements in computation and transmission load.
III. C ONCLUSIONS
6G research has already started around the world address-
ing innovative solutions to offer at the horizon of 2030 new
services. In this paper we first presented the current vision in
the research community on what should be in 6G that is not
in 5G from an air interface perspective.
The next generation of wireless systems will transform the
5G service-oriented networks into user and machine ad-hoc dy-
namic (re)configuration of network slices. This will be enabled
by software-defined end-to-end solutions from the core to the
radio access network, including the air interface as well as the
RF and antenna systems. This paper presents our perspective of
the 6G air interface and raise the concept of software defined
artificial intelligence and air interface as a framework of the
next generation of air interface. We expect to tune the air
interface according to each user requirements by taking into
account, the expected quality of service, the localization, the
propagation environment, the available bandwidth and carrier
frequency and so on. It includes the use of new spectrum
opportunities such as sub-THz and optical spectrum. We also
envisage artificial intelligence (learning scheme), distributed
computing and caching techniques to play a central role in
6G.
R EFERENCES
[1] Q. Sun, C. I, S. Han, Z. Xu, and Z. Pan, “Software defined air
interface: a framework of 5G air interface,” in 2015 IEEE Wireless
Communications and Networking Conference Workshops (WCNCW),
March 2015, pp. 6–11.
[2] E. Calvanese-Strinati and et al, “5GCHAMPION Disruptive 5G Tech-
nologies for RollOut in 2018,” in ETRI Journal 40.1, 2018.
[3] Netword2020 association, “White Paper for Research Beyond 5G,” in
ETRI Journal 40.1, 2015.
[4] E. Calvanese-Strinati, S. Barbarossa, J.-L. Gonzalez-Jimenez, L. Maret,
D. Kténas, N. Cassiau, and C. Dehos, “6G: The Next Frontier,” in arXiv
preprint arXiv:1901.03239, 2019.
[5] D. Demmer, R. Gerzaguet, J.-B. Doré, and D. Le-Ruyet, “Analytical
study of 5G NR eMBB co-existence,” in 2018 25th International
Conference on Telecommunications (ICT), June 2018, pp. 186–190.
[6] N. di Pietro, M. Merluzzi, E. Calvanese-Strinati, and S. Barbarossa,
“Resilient design of 5G mobile-edge computing over intermittent
mmWave links,” in submitted to IEEE Trans. Mobile Comput. [Online]
Available: http://arxiv.org/abs/1901.01894, 2019.
[7] D. Demmer, , R. Zakaria, J.-B. Doré, R. Gerzaguet, and D. Le-Ruyet,
“Filter-bank OFDM transceivers for 5G and beyond,” in Asilomar
Conference on Signals, Systems, and Computers, 2018, Nov 2018.
[8] J. Y.-K. T. Levanen, M. Renfors, M. Valkama, and K. Pajukoski,
“FFT-Domain Signal Processing for Spectrally-Enhanced CP-OFDM
Waveforms in 5G New Radio,” in Asilomar Conference on Signals,
Systems, and Computers, 2018, Nov 2018.
[9] A. F. Molisch, V. V. Ratnam, S. Han, Z. Li, S. L. H. Nguyen, L. Li, and
K. Haneda, “Hybrid beamforming for massive mimo: A survey,” IEEE
Communications Magazine, vol. 55, no. 9, pp. 134–141, Sep. 2017.
[10] P.-R. Chang, W.-H. Yang, and K.-K. Chan, “A neural network approach
to mvdr beamforming problem,” IEEE Transactions on Antennas and
Propagation, vol. 40, no. 3, pp. 313–322, March 1992.
[11] S. Navabi, C. Wang, O. Y. Bursalioglu, and H. Papadopoulos, “Predict-
ing Wireless Channel Features Using Neural Networks,” in 2018 IEEE
International Conference on Communications (ICC), June 2018.
[12] J. I. AlHajri, N. T. Ali, and R. M. Shubair, “A Machine Learning
Approach for the Classification of Indoor Environments Using RF
Signatures,” in IEEE Global Conference on Signal and Information
Processing (GlobalSIP 2018), Nov 2018.